VoIP & Communications
Is VoIP reliable enough for business calls? What actually determines call quality
VoIP is reliable for business calls when the network under it is right. Learn the latency, jitter and packet loss limits, QoS, 4G failover and the common myths.
By Dig IT SolutionsUpdated 8 September 20267 min read
Short answer
Yes, VoIP is reliable enough for business calls, provided the network underneath it is designed for voice. Call quality depends on latency under about 150 ms, jitter under about 30 ms and packet loss under 1 percent, which a business-grade connection with quality of service delivers. With 4G backup and hosted failover, VoIP is more resilient than a PBX.
The question every business asks before moving off landlines is whether the calls will sound as good and stay up as reliably. The honest answer is that VoIP reliability is a network question, not a telephony one. Get the connection, the router and the cabling right and hosted VoIP is clearer and more resilient than the PBX it replaces. Get them wrong and no platform will save you. This article sets out what determines call quality, how to design for it, and why the old objections no longer hold.
What VoIP needs from your network
Voice is converted into small packets and sent across the internet, then reassembled at the other end in real time. Three measurements decide whether that sounds natural:
| Measure | What it is | Target for good calls | What you hear when it is exceeded |
|---|---|---|---|
| Latency | Delay from speaker to listener | Under about 150 ms one way | People talking over each other, awkward pauses |
| Jitter | Variation in that delay | Under about 30 ms | Robotic or warbling voice |
| Packet loss | Packets that never arrive | Under 1 percent | Gaps, dropped words, one-way audio |
Bandwidth, the figure everyone worries about, is rarely the constraint. A call uses around 100 kbps in each direction, so 20 concurrent calls need about 2 Mbps. What breaks calls is a saturated connection where voice queues behind a cloud backup or a video meeting, a router that treats all traffic equally, a phone on Wi-Fi, or a cable that was never tested.
How to design for reliable calls
Business-grade connectivity. A fibre (FTTP) or SoGEA connection with a business service level, rather than a consumer broadband package shared with the flat upstairs. Where calls are critical, a second connection from a different provider or a 4G or 5G backup on the router.
Quality of service (QoS). The router and switches must be configured to prioritise voice packets over everything else. This is the single most common omission in self-installed VoIP and the most common cause of "the phones are rubbish" complaints.
A voice VLAN. Desk phones on their own network segment, separated from PCs, guest Wi-Fi, CCTV and printers.
Wired phones, tested cabling. Desk phones on Ethernet, not Wi-Fi. Cable runs certified, because a marginal cable produces exactly the symptoms people blame on VoIP. Our structured cabling work often precedes a phone rollout for this reason.
Power protection. A UPS on the comms cabinet so a short power cut does not drop every call in the building.
Monitoring. Call quality statistics from the platform reviewed regularly, so a deteriorating connection is spotted before users start complaining.
This is the standard scope of Dig IT's VoIP and business phone systems service and it is why we assess the network before we quote for phones.
Failover: what happens when something breaks
Reliability is also about what happens when, not if, a component fails. Hosted VoIP has an advantage here that a PBX in a cabinet cannot match: the platform is in the provider's data centres, with redundant power and connectivity, so a fault in your office affects your office only.
- Broadband fault: a 4G or 5G backup on the router keeps desk phones and PCs working. Without one, calls divert to mobile apps automatically.
- Power cut: UPS covers short cuts. Longer ones, staff answer on the mobile app.
- Office closed: everyone logs in from home on the same numbers. See VoIP and business continuity.
- Handset failure: swap the phone, log in, nothing else changes.
- Platform maintenance: hosted providers run across multiple data centres and publish their availability commitments.
The rule is to configure and test the diversion before go-live. A failover that has never been tried is not a failover.
Why inconsistent quality costs more than you think
Customers do not separate a bad line from a bad business. A crackly call, a dropped call mid-negotiation or a voice that keeps breaking up reads as disorganisation, and people rarely complain about it. They just stop calling. For solicitors, financial advisers, estate agents and IT providers, where the phone call is the service, the reputational cost of poor audio is real and invisible.
Inconsistency is worse than consistently mediocre. If one office sounds fine and another does not, or one home worker is clear and another keeps dropping, the business appears unreliable even though most calls are fine. That is why remote staff need the same standards as the office: a headset, a decent connection and a quick call-quality test during onboarding.
The common objections, answered
"VoIP call quality is worse than a landline." It was, in the early 2000s on slow, contended broadband. On a modern business connection with QoS, HD voice codecs deliver noticeably clearer audio than an analogue line. Where quality is poor, the network is the reason.
"VoIP dies when the internet goes down." Desk phones in the office do, unless there is a 4G backup. The platform does not, and calls divert to mobiles. A PBX with the lines cut has no equivalent.
"It is not secure." Business platforms encrypt calls and signalling, support multi-factor authentication and monitor for fraud. The risks are configuration risks: default passwords, open international dialling, too many admins. Those are fixable, and an analogue line offered no controls at all.
"It is only for tech companies." Most of the businesses moving to VoIP now are trades, professional services, hire companies and clinics. The admin is a web portal. The training is a one-page guide.
"It is expensive." Per-user subscriptions typically cost less than line rental plus maintenance on a PBX, and inter-site calls are free. See what VoIP costs in the UK.
"We will lose our numbers." Number porting is standard and regulated. You keep them.
"Switching is disruptive." Run in parallel for a week or two, port on an agreed date, and callers notice nothing. The process is in how to switch from a landline to VoIP.
"It is the same as WhatsApp or Skype." Consumer apps use similar technology but offer no queues, no routing, no reporting, no support and no business numbers.
A pre-go-live reliability checklist
Before any phones go live, work through this list. Every item is something we have seen cause a "VoIP is unreliable" complaint that turned out to be nothing to do with VoIP.
- Connection tested at the busiest hour of the day for latency, jitter and packet loss, not just download speed.
- Router confirmed to support QoS, with voice prioritised and the setting actually enabled.
- Desk phones on a separate voice VLAN and on Ethernet, with PoE switch capacity checked.
- Cabling to every phone position tested and certified.
- 4G or 5G backup installed and a failover actually forced to prove it works.
- UPS on the comms cabinet with enough runtime for a typical local power cut.
- Divert rules to mobile apps configured and tested by unplugging the broadband.
- Registered emergency address correct for every site.
- Default passwords changed, multi-factor authentication on, international dialling restricted, admin rights limited to named people.
- Home workers' call quality checked from their actual home connection with the headset they will use.
- Call quality dashboard reviewed weekly for the first month.
A business that ticks every item will find the phones are the most dependable system it runs. A business that skips the first two will be back on the phone to its provider within the week.
When VoIP is genuinely the wrong answer
There are a few cases where hosted VoIP alone is not enough. A remote site with no fibre, no usable 4G and a single poor copper connection will struggle until connectivity improves. Lift emergency phones, monitored alarm panels and some telecare devices need their own IP or mobile-network solution rather than a VoIP handset. A business that cannot tolerate any dependency on mains power or broadband at a site needs a mobile fallback designed in from the start. None of these is an argument for staying on analogue lines, which are being withdrawn in any case, but they are reasons to plan rather than assume.
Emergency calls and compliance
UK VoIP providers must provide access to emergency services and hold a registered address for each number. Keep those addresses correct for every site and update them when people move. Ofcom's guidance on the move to digital landlines also expects providers to offer a solution, such as battery back-up, for customers who depend on a landline to reach emergency services during a power cut. For businesses with lifts, alarms or lone workers, that is a design point rather than a footnote.
What to do next
If you are hesitating over VoIP because of reliability, the useful first step is a network assessment rather than a platform demo. Dig IT tests connectivity, router configuration and cabling, and tells you what would need to change before the phones go in. Talk to an engineer and we will give you a straight answer on whether your network is ready.